Título:	Um novo modelo de regressão Birnbaum–Saunders generalizado : uma abordagem via GAMLSS
Autor(es):	Milhomem, Matheus Braga
Orientador(es):	Ribeiro, Terezinha Késsia de Assis
Assunto:	Modelo de regressão Simulação de Monte Carlo Modelagem de dados Modelo aditivo generalizado
Data de publicação:	25-Jun-2026
Data de defesa:	1-Jan-2026
Referência:	MILHOMEM, Matheus Braga. 2026. 73 f., il. Dissertação (Mestrado em Estatística) — Universidade de Brasília, Brasília, 2026.
Abstract:	The Birnbaum–Saunders (BS) distribution is a probabilistic model that has gained increasing prominence in the literature for modeling positive and asymmetric continuous data. Several authors have proposed extensions of this probabilistic model, and some have developed regression frameworks for modeling its parameters. In this study, a new class of regression models is proposed for the generalization of the BS distribution introduced by Owen (2006). The Generalized Birnbaum–Saunders (GBS) distribution considered herein has received relatively limited attention in the literature, despite reinforcing the physical justification of the originally derived model. The specification of the proposed regression model is analogous to that of the Generalized Additive Models for Location, Scale, and Shape (GAMLSS), thereby providing a more flexible framework for data fitting. Parameter estimation is carried out via the maximum likelihood method, with the inferential procedure constructed from the formulation of the likelihood function and the derivation of the associated score vectors for the model parameters. The GBS model was implemented computationally in a novel manner within the R environment through the gamlss package, enabling the use of a comprehensive set of tools for fitting and assessing the proposed model. Monte Carlo simulation studies were conducted to evaluate the behavior and performance of the estimation procedure under the proposed GBS regression model. The results indicate good performance of the maximum likelihood estimator, as evidenced by decreasing bias and increasing efficiency as the sample size increases. Finally, the applicability of the proposed model is illustrated and discussed through applications to real data, empirically demonstrating scenarios in which the GBS model provides a better fit to the data when compared with the original BS model.
Unidade Acadêmica:	Instituto de Ciências Exatas (IE) Departamento de Estatística (IE EST)
Informações adicionais:	Dissertação (mestrado) — Universidade de Brasília, Instituto de Ciências Exatas, Departamento de Estatística, 2026.
Programa de pós-graduação:	Programa de Pós-Graduação em Estatística
Agência financiadora:	Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES)
Aparece nas coleções:	Teses, dissertações e produtos pós-doutorado

Os itens no repositório estão protegidos por copyright, com todos os direitos reservados, salvo quando é indicado o contrário.